Stable Diffusion WebUI Forge - Neo

[ Neo | Classic ]

UI

Stable Diffusion WebUI Forge is a platform on top of the original Stable Diffusion WebUI by AUTOMATIC1111, to make development easier, optimize resource management, speed up inference, and study experimental features.
The name "Forge" is inspired by "Minecraft Forge". This project aims to become the Forge of Stable Diffusion WebUI.

- lllyasviel
(paraphrased)


"**Neo**" mainly serves as an continuation for the "`latest`" version of Forge, which was built on [Gradio](https://github.com/gradio-app/gradio) `4.40.0` before lllyasviel became too busy... Additionally, this fork is focused on optimization and usability, with the main goal of being able to run the latest popular models via an easy-to-use GUI. > [!Tip] > [How to Install](#installation)
## Features [Jul.] > Most base features of the original [Automatic1111 Webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui) should still function #### New Features - [X] Support [Krea 2](https://huggingface.co/krea/Krea-2-Turbo) - `Turbo` / `Raw` - [X] Support [Anima](https://huggingface.co/circlestone-labs/Anima) - [X] Support **Anima Edit** - require specific [LoRA](https://civitai.com/models/2650553/anima-edit) - enable in **Settings/Stable Diffusion** - [X] Support [Flux.2-Klein](https://huggingface.co/black-forest-labs/FLUX.2-klein-4B) - `4B` / `9B` (**not** `FLUX.2-Dev`) > [!Important] > To use `Flux.2-Klein` for regular `img2img`, toggle the functionality in **Settings/Stable Diffusion** - [X] Support [Ernie-Image](https://huggingface.co/baidu/ERNIE-Image) - `ernie-image` / `ernie-image-turbo` - [X] Support [PiD 1.0](https://huggingface.co/nvidia/PiD) - `sdxl` / `qwen` / `flux1` / `flux2` (**not** `PixelDiT`) - use `PiD Integrated` to automatically upscale after generation - [X] Support [Z-Image](https://huggingface.co/Tongyi-MAI/Z-Image) - `z-image` / `z-image-turbo` - [X] Support [Wan 2.2](https://github.com/Wan-Video/Wan2.2) - `14B` (**not** `5B`) - use `Refiner` to achieve **High Noise** / **Low Noise** switching - enable `Refiner` in **Settings/Refiner** > [!Important] > To export a video, you need to have **[FFmpeg](https://ffmpeg.org/)** installed - [X] Support [Mugen](https://huggingface.co/CabalResearch/Mugen) - display the `Shift` slider for `xl` preset in **Settings/Presets/XL** - [X] Support advanced **SDXL** models > [!Note] > - **v-prediction:** `state_dict` must include "`v_pred`" > - **Zero Terminal SNR:** `state_dict` must include "`ztsnr`" > - **Rectified Flow:** the model must include "`rectified`" in its path *(**e.g.** file name or folder name)* - [X] Support [Qwen-Image](https://huggingface.co/Qwen/Qwen-Image) / [Qwen-Image-Edit](https://huggingface.co/Qwen/Qwen-Image-Edit-2509) > [!Note] > To be detected as an **Edit** model, the model must include "`qwen`" and "`edit`" in its path *(**e.g.** file name or folder name)* - [X] Support [Flux Kontext](https://huggingface.co/black-forest-labs/FLUX.1-Kontext-dev) > [!Note] > To be detected as a **Kontext** model, the model must include "`kontext`" in its path *(**e.g.** file name or folder name)* - Implement `ImageStitch Integrated` - [X] support Multi-Image Inputs for `flux.2-klein` / `flux-kontext` / `qwen-image-edit` - [X] support FirstLastFrameToVideo for `wan 2.2` - [X] Support [Nunchaku](https://github.com/nunchaku-tech/nunchaku) (`SVDQ`) Models - `flux-dev`, `flux-krea`, `flux-kontext`, `qwen-image`, `qwen-image-edit`, `z-image-turbo` - only `Flux` and `Qwen` support LoRA currently - see [Commandline](#by-neo) - [X] Support [Lumina-Image-2.0](https://huggingface.co/Alpha-VLLM/Lumina-Image-2.0) - `Neta-Lumina` / `NetaYume-Lumina` - [X] Support [Chroma1-HD](https://huggingface.co/lodestones/Chroma1-HD) - [X] Support **MixedPrecision** Models - `fp4mixed` / `fp8mixed` / `mxfp8` / `nvfp4` / `fp8_scaled` / `int8_convrot` / `convrot_w4a4` - [X] Support [Flux.2-Small-Decoder](https://huggingface.co/black-forest-labs/FLUX.2-small-decoder/blob/main/full_encoder_small_decoder.safetensors) & [Qwen2D VAE](https://huggingface.co/Anzhc/Qwen2D-VAE/blob/main/Qwen2D_VAE.safetensors)
> [!Tip] > Check out [Download Models](https://github.com/Haoming02/sd-webui-forge-classic/wiki/Download-Models) for where to get each model and the accompanying modules > [!Tip] > Check out [Inference References](https://github.com/Haoming02/sd-webui-forge-classic/wiki/Inference-References) for how to use each model and the recommended parameters
- [X] Rewrite Preset System - now save the checkpoint/module selection and parameters per each Preset > [!Note] > This overrides the `UI Defaults` for the controlled parameters
- [X] Enforce Resolution Steps - dimensions must be multiples of `64` by default - adjust in **Settings/System** - [X] Support [uv](https://github.com/astral-sh/uv) package manager - drastically speed up installation - require **manually** installing [uv](https://github.com/astral-sh/uv/releases) - see [Commandline](#by-neo) - [X] Support [SageAttention](https://github.com/thu-ml/SageAttention), [FlashAttention](https://github.com/Dao-AILab/flash-attention), `fp16_accumulation`, `torch._scaled_mm` - see [Commandline](#by-neo) - [X] Implement Triton Kernel for `matmul` in `torch.int8` - speed up inference after quantization - enable by selecting `int8` in the `Diffusion in Low Bits` - [X] Implement [Radial Attention](https://github.com/mit-han-lab/radial-attention) - speed up `Wan 2.2` - require **manually** installing [SpargeAttn](https://github.com/thu-ml/SpargeAttn) - [X] Implement fast `state_dict` switching for Refiner - enable in **Settings/Refiner** - [X] Implement RescaleCFG - reduce burnt colors; mainly for `v-pred` checkpoints - enable in **Settings/UI Alternatives** - [X] Implement MaHiRo - alternative CFG calculation; improve prompt adherence - enable in **Settings/UI Alternatives** - [X] Implement [Spectrum](https://github.com/hanjq17/Spectrum) - training-free acceleration for all models - [X] Implement [Epsilon Scaling](https://github.com/comfyanonymous/ComfyUI/pull/10132) - enable in **Settings/Stable Diffusion** - [X] Implement `torch.compile` - speed up inference after compilation - [X] Implement alternative Prompt Box layouts - [X] Implement tiled `Conv2d` for VAE - reduce memory usage; reduce speed - see [Commandline](#by-neo) - [X] Implement full precision calculation for `Mask blur` blending - enable in **Settings/img2img** - [X] Support TAESD / TAEHV live preview for all models - [X] Support video previews for ExtraNetworks - [X] Support loading upscalers in `half` precision - speed up; reduce quality - enable in **Settings/Upscaling** - [X] Support running tile composition on GPU - enable in **Settings/Upscaling** - [X] Support (short) videos in **Extras** tab - [X] Add support for `.avif`, `.heif`, and `.jxl` image formats - [X] Automatically determine the optimal row count for `X/Y/Z Plot` - [X] Update **LLLite** Controlnet - [SDXL](https://huggingface.co/kohya-ss/controlnet-lllite/tree/main) / [Anima](https://huggingface.co/kohya-ss/Anima-LLLite/tree/main) - [X] Support **Union** Controlnet - [SDXL](https://huggingface.co/xinsir/controlnet-union-sdxl-1.0) / [Chenkin](https://civitai.com/models/2527960/chenkin-unicontrol-xl) - [X] Support **Region** Controlnet - [Anima](https://huggingface.co/Sen-sou/Anima-LLLite-Regional-Controlnet) - use the **Region** mode to create color masks #### Removed Features - [X] SD2 - [X] SD3 - [X] Forge Spaces - [X] Hypernetworks - [X] CLIP Interrogator - [X] Deepbooru Interrogator - [X] Textual Inversion Training - [X] Some built-in Extensions - [X] Some built-in Scripts - [X] Some Samplers & Schedulers - [X] Some Compatibility Settings - [X] Stealth Infotext #### Optimizations - [X] **[Comfy]** Rewrite the Backend *(`memory_management.py`, `ModelPatcher`, `attention.py`, etc.)* - [X] No longer `git` `clone` any repository on fresh install - [X] No longer install `open-clip` - [X] Fix memory leak when switching checkpoints - [X] Restore the ability to drag-and-drop images onto `gr.Image` that already contains image - [X] Speed up launch time - [X] Improve timer logs - [X] Remove unused `cmd_args` - [X] Remove unused `args_parser` - [X] Remove unused `shared_options` - [X] Remove legacy codes - [X] Fix some typos - [X] Fix automatic `Tiled VAE` fallback - [X] Fix `Tiling` for SD1 and SDXL - [X] Pad conditioning for SDXL - [X] Remove duplicated upscaler codes - [X] Update [spandrel](https://github.com/chaiNNer-org/spandrel) - support new upscaler architectures > [!Important] > Put every upscaler (`.pth` / `.safetensors`) inside the `ESRGAN` folder > [!Tip] > Check out [OpenModelDB](https://openmodeldb.info/) for where to get upscalers - [X] Improve `ForgeCanvas` - brush adjustments - customization - deobfuscate - eraser - hotkeys - mobile friendly - [X] Optimize upscaler logics - [X] Optimize certain operations in `Spandrel` - [X] Optimize certain operations for `VAE` - [X] Speed up model loading - [X] Improve color correction - [X] Improve memory management - [X] Unload all models on `OutOfMemory` - [X] Update the implementation for `X/Y/Z Plot` - [X] Update the implementation for `Soft Inpainting` - [X] Update the implementation for `MultiDiffusion` - [X] Update the implementation for `uni_pc` and `LCM` samplers - [X] Update the implementation of LoRAs - [X] Revamp settings - improve formatting - update descriptions - [X] Check for Extension updates in parallel - [X] Move `embeddings` folder into `models` folder - [X] Infotext Rewrite - allow switching Models and Modules - save `emphasis` properly - correct default values - [X] ControlNet Rewrite - change Units to `gr.Tab` - improve `masks` & `buttons` - remove multi-inputs, as they are "[misleading](https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/932)" - [X] Checkpoint Merger Rewrite - update ui - support modern models - [X] Disable Refiner by default - enable again in **Settings/Refiner** - [X] No longer install `bitsandbytes` by default - see [Commandline](#by-neo) - [X] Improved non-Nvidia support - [X] Lint & Format - [X] Update `Pillow` - faster image processing - [X] Update `protobuf` - faster `insightface` loading - [X] Update to latest PyTorch - `torch==2.11.0+cu130` > [!Note] > If your GPU does not support the latest PyTorch, manually [install](https://github.com/Haoming02/sd-webui-forge-classic/wiki/Extra-Installations#older-pytorch) older version of PyTorch - [X] Update some packages to newer versions - [X] Update recommended Python to `3.13.12` - [X] many more... :tm:
## Commandline > These flags can be added after the `set COMMANDLINE_ARGS=` line in the `webui-user.bat` *(in the same line ; separate each flag with space)* > [!Tip] > Use `python launch.py --help` to see all available flags - `--xformers`: Install the `xformers` package to speed up generation > [!Warning] > `xformers` does **not** support `RTX 50s` - `--port`: Specify a server port to use - defaults to `7860` - `--api`: Enable [API](https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/API) access #### by. Neo - `--cuda-malloc`: Improve memory allocation - `--cuda-stream`: Enable async weight offloading - `--pin-shared-memory`: Improve RAM utilization - `--expandable-segments`: Enable experimental PyTorch allocator *(may prevent `OutOfMemory` errors on certain platforms)*
- `--uv`: Replace the `python -m pip` calls with `uv pip` to massively speed up package installation - requires **uv** to be installed first *(see [Extra Installations](https://github.com/Haoming02/sd-webui-forge-classic/wiki/Extra-Installations))* - `--uv-symlink`: Same as above; but additionally pass `--link-mode symlink` to the commands - significantly reduces installation size (`~7 GB` to `~100 MB`) - `--uv-local-cache`: Same as above; but additionally set `UV_CACHE_DIR` to a `.uv-cache` folder within WebUI directory - speed up installation on non-default drive *(**i.e.** not `C:` on Windows)* - allow clean uninstallation by simply deleting the WebUI directory > [!Important] > `symlink` means it will directly access the packages from the cache folder instead of copying the packages over ; refrain from clearing the cache when using this option
- `--model-ref`: Points to a central `models` folder that contains all your models - said folder should contain subfolders like `Stable-diffusion`, `Lora`, `VAE`, `ESRGAN`, etc. > [!Important] > This simply **replaces** the `models` folder rather than adding on top of it
- `--forge-ref-a1111-home`: Point to an Automatic1111 installation to load its `models` folders - **i.e.** `Stable-diffusion`, `text_encoder`, etc. - `--forge-ref-comfy-home`: Point to a ComfyUI installation to load its `models` folders - **i.e.** `diffusion_models`, `clip`, etc. - `--forge-ref-comfy-yaml`: Point to the ComfyUI `extra_model_paths.yaml` to load its configurations - **i.e.** `base_path`, `checkpoints`, etc.
- `--sage`: Install the `sageattention` package to speed up generation - will also attempt to install `triton` automatically - `--flash`: Install the `flash_attn` package to speed up generation - `--nunchaku`: Install the `nunchaku` package to inference SVDQ models - `--bnb`: Install the `bitsandbytes` package to do low-bits (`nf4`) inference - `--onnxruntime-gpu`: Install the `onnxruntime` with the latest GPU support
- `--fast-fp8`: Use the `torch._scaled_mm` function when the model type is `float8_e4m3fn` - `--fast-fp16`: Enable the `allow_fp16_accumulation` option - `--autotune`: Enable the `torch.backends.cudnn.benchmark` option - this is slower in my experience... - `--tiled-conv2d`: Replace `Conv2d` ops with tiled variants - has greater reduction for **SD1** and **SDXL** VAE; less for **Wan** VAE - `64` / `128` / `256` / `512`
## Installation 0. Install **[git](https://git-scm.com/downloads)** 1. Clone the Repo ```bash git clone https://github.com/Haoming02/sd-webui-forge-classic sd-webui-forge-neo --branch neo ``` 2. Setup Python
Recommended Method - Install **[uv](https://github.com/astral-sh/uv#installation)** - Set up **venv** ```bash cd sd-webui-forge-neo uv venv venv --python 3.13 --seed ``` - Add the `--uv` flag to `webui-user.bat`

Deprecated Method - Install **[Python 3.13.12](https://www.python.org/downloads/release/python-31312/)** - Remember to enable `Add Python to PATH`

3. **(Optional)** Configure [Commandline](#commandline) 4. Launch the WebUI via `webui-user.bat` 5. During the first launch, it will automatically install all the requirements 6. Once the installation is finished, the WebUI will start in a browser automatically
> [!Tip] > - For **AMD**, refer to CS1o 's Guide > - For **Linux** and **macOS**, refer to [Wiki](https://github.com/Haoming02/sd-webui-forge-classic/wiki/Unix) > - For **Docker** (`Nvidia`), refer to [Docker](docker/)
> [!Tip] > Check out [Extra Installations](https://github.com/Haoming02/sd-webui-forge-classic/wiki/Extra-Installations) for how to install `git`, `uv`, and `FFmpeg`
## Attention Functions > [!Important] > The `--xformers`, `--flash`, and `--sage` args are only responsible for installing the packages, **not** whether its respective attention is used *(this also means you can remove them once the packages are successfully installed)* > [!Caution] > Do **not** just blindly install all of them
> Nowadays the native PyTorch `scaled_dot_product_attention` is usually as fast, and also more stable **Forge Neo** tries to import the packages and automatically choose the first available attention function in the following order: 1. `SageAttention` 2. `FlashAttention` 3. `xformers` 4. `PyTorch` 5. `Basic` > [!Note] > To skip a specific attention, add the respective disable arg such as `--disable-sage`
## Issues & Requests - **Issues** about removed features will simply be ignored - **Issues** that is obviously user-error will simply be ignored - **Issues** regarding **AMD** GPU will simply be ignored - **Issues** running non-official models will simply be ignored - do not just randomly download every single finetune/quant you find - **Issues** about 3rd-party Extensions will simply be ignored - extension should support the UI, not the other way around - **Issues** caused by [StabilityMatrix](https://github.com/LykosAI/StabilityMatrix) will simply be ignored - only open an Issue if you can reproduce it on a clean install following the official [Installation](#installation) instruction > [!Caution] > - If you post **NSFW** images/videos, you will immediately be banned > - the sole discretion is on me ; if you are unsure, just generate `cats` and `dogs`...
> [!Tip] > Check out the [Wiki](https://github.com/Haoming02/sd-webui-forge-classic/wiki) & [FAQ](https://github.com/Haoming02/sd-webui-forge-classic/issues/414)

Special thanks to AUTOMATIC1111, lllyasviel, and comfyanonymous, kijai, city96,
along with the rest of the contributors,
for their invaluable efforts in the open-source image generation community


Buy me a Coffee ☕~
PayPal me 💳~